Paper Title: Predicting and Identifying Potato Blight Disease through Deep Learning with Fuzzy Approaches

Author:

Debraj Roy¹, Biswapati Jana² and Nguyen Kim Sao³
¹Department of Computer Science, Tamralipta Mahavidyalaya, Tamluk, Purba Medinipur,721636, West Bengal, India, debraj545roy@gmail.com
²Department of Computer Science, Vidyasagar University, Paschim Medinipur, West Bengal, India, biswapatijana@gmail.com
³Department of Computer Science, University of Transport and Communication, Đ. Cầu Giấy, Láng Thượng, Đống Đa, Hà Nội, Vietnamsaonkoliver@gmail.com
AIJITR, Volume-1, Issue-I, September- October, 2024, PP.69-77.
Revised and accepted on October 16, 2024

Abstract:

Two diseases, early and late blight, put great danger on the potato crop and cause farmers to lose money. Farmers could immediately react to crops that these diseases have impacted be- cause of their early and automatic identification. The deep learning methodology is extensively discussed in the literature and provides various possibilities for diagnosing crop blight diseases. This article uses deep learning to identify the disease in crops. Here, machine learning-based image processing methods have been taken into consideration. We have employed VGG16, ResNet50, and ResNet152V2 models and the fuzzy technique based on deep learning to identify the blight disease stage in potato leaf images. To make pixel-by-pixel forecasts in pictures of specific leaves, the initial instance of the state-of-the-art DeepLabV3+ linguistic segmentation framework, constructed from ResNet152V2, is trained. Then, the features that were extracted, such as ROI and POI. Second, an uncertain rulebased method is created for every characteristic to estimate the severity of disease harm. In the fuzzy logic system, suitable membership functions for the inputs and outputs are also considered for fuzzification and defuzzification processes. Last, potato leaves are labeled Healthy, Mild, Medium, and Severe. However, utilizing ResNet152V2, we were able to reach 98% accuracy, the best result among these techniques. 

Keywords:Deep learning; Fuzzy Inference System; Blight infection; Potato (Solanum tuberosum L.), Deep learning, Convolutional Neural Networks (CNNs)

Doi Link –

Review By – Prof. Dr. Sandeep Poddar, Dr. Shivalika Sarkar and Dr. Oyyappan Duraipandi,